Commits · b41aeb3468fecb33b7934fc7ca5406bdba84dace · OpenDAS / vllm_cscc

24 Dec, 2025 5 commits
- [Bugfix][ROCm] Fix load issue on deepseek quark quantization when shared expert enabled (#31261) · b41aeb34
  Pleaplusone authored Dec 24, 2025
```
Signed-off-by: ganyi <ygan@amd.com>
```
  b41aeb34
- [Qwen3-Omni] fixed _get_feat_extract_output_lengths function (#31007) · bb24592d
  Xiong Wang authored Dec 24, 2025
```
Signed-off-by: Xiong Wang <wangxiongts@163.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
```
  bb24592d
- [ROCm][CI][Bugfix] Fix Siglip2 rotary embedding dispatch and InternVL video test tolerance (#31235) · e42894f5
  Andreas Karatzas authored Dec 23, 2025
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  e42894f5
- [Bug] Fix `Number of dimensions of tensors must match.` for Deepseek V3.2 (#31160) · 76e6a951
  Wentao Ye authored Dec 23, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  76e6a951
- [Bugfix] Enable `dynamic_dims` for different embeds shape (#31223) · dd424571
  Cyrus Leung authored Dec 24, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  dd424571
23 Dec, 2025 4 commits
- adapt voxtral (#31095) · 3faa8bee
  Patrick von Platen authored Dec 23, 2025
```
Signed-off-by: Patrick von Platen <patrick.v.platen@gmail.com>
```
  3faa8bee
- Add util function for checking nesting of rope parameters (#31146) · b10d47e0
  Harry Mellor authored Dec 23, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  b10d47e0
- [Frontend] Support using chat template as custom score template for reranking models (#30550) · 23daef54
  Jakub Zakrzewski authored Dec 23, 2025
```
Signed-off-by: Jakub Zakrzewski <jzakrzewski@nvidia.com>
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
Signed-off-by: wang.yuqi <noooop@126.com>
Co-authored-by: wang.yuqi <yuqi.wang@daocloud.io>
```
  23daef54
- [Bugfix] Fix Jais2ForCausalLM (#31198) · 6b16fff0
  Jee Jee Li authored Dec 23, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  6b16fff0
22 Dec, 2025 3 commits
- [BugFix] skip language model in Encoder (#30242) · 8f8f469b
  dengyunyang authored Dec 22, 2025
```
Signed-off-by: dengyunyang <584797741@qq.com>
```
  8f8f469b
- [Model] Fix bagel failed to run (#31132) · 256a33ec
  Li Wang authored Dec 22, 2025
```
Signed-off-by: wangli <wangli858794774@gmail.com>
```
  256a33ec
- [Misc] Fix spelling typos in model comments (#31117) · 14c3e6ad
  Kevin McKay authored Dec 21, 2025
```
Signed-off-by: c0de128 <kevin.mckay@outlook.com>
```
  14c3e6ad
20 Dec, 2025 2 commits

[MoE Refactor][5/N] Isolate zero expert to LongCatFlash (#28891) · 54c89243

baonudesifeizhai authored Dec 20, 2025

Signed-off-by: baonudesifeizhai <85092850+baonudesifeizhai@users.noreply.github.com>
Signed-off-by: Dongjie Zou <85092850+baonudesifeizhai@users.noreply.github.com>
Signed-off-by: baonudesifeizhai <baonudesifeizhai@gmail.com>
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robertgshaw2@gmail.com>

54c89243

GLM-4.7 Tool Parser and Doc Update (#30876) · 8a7a4143
Yuxuan Zhang authored Dec 20, 2025
```
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
```
8a7a4143

19 Dec, 2025 3 commits

[Bug] Fix `error 'Dynamo failed to run FX node with fake tensors` for Deepseek V3.2 (#31046) · 4cf94298
Wentao Ye authored Dec 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
4cf94298

[Model] Add MiMo-V2-Flash support (#30836) · 969bbc7c

Zhonghua Deng authored Dec 20, 2025


Signed-off-by: Abatom <abzhonghua@gmail.com>
Signed-off-by: Jumiar <liuanqim10@126.com>
Signed-off-by: Zyann7 <zyann7@outlook.com>
Co-authored-by: Jumiar <liuanqim10@126.com>
Co-authored-by: Zyann7 <zyann7@outlook.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>

969bbc7c

[ROCm][CI][Bugfix] Multi-Modal Model Support Fixes and Attention Backend Improvements (#30270) · 7b43db21
Andreas Karatzas authored Dec 18, 2025
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
7b43db21

18 Dec, 2025 5 commits
- [MM Encoder]: Migrate legacy ViT `MultiHeadAttention` to new... · 700a5ad6
  Isotr0py authored Dec 19, 2025
```
[MM Encoder]: Migrate legacy ViT `MultiHeadAttention` to new `MMEncoderAttention` interface (#30684)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  700a5ad6
- adds jais 2 support (#30188) · 28d15ab5
  sarathc-cerebras authored Dec 18, 2025
```
Signed-off-by: sarathc-cerebras <sarath.chandran@cerebras.net>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
```
  28d15ab5
- [refactor] Add prefix support to embed_tokens in DeepSeek MTP (#30788) · b166ef20
  zzhxxx authored Dec 18, 2025
```
Signed-off-by: zzhx1 <zzh_201018@outlook.com>
```
  b166ef20
- [Chore] Remove v0 dead code for Qwen2.5-omni (#30883) · 6fe58876
  Isotr0py authored Dec 18, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  6fe58876
- [v1] Add PrefixLM support to TritonAttention backend (#30386) · 74a1ac38
  Isotr0py authored Dec 18, 2025
  
  74a1ac38
17 Dec, 2025 4 commits
- [Model] Gemma3: Support untied word embeddings (#30827) · 196cdc32
  KimHyemin authored Dec 18, 2025
```
Signed-off-by: www-spam <panmahm@naver.com>
```
  196cdc32
- [Bugfix] deepseek-V3.2 self.weights_proj has no bias (#30841) · 84896fda
  baoqian426 authored Dec 17, 2025
```
Signed-off-by: baoqian <1354987947@qq.com>
Signed-off-by: baoqian426 <1354987947@qq.com>
```
  84896fda
- [Mamba] Removed disable cascade attn in MambaModelConfig (#30712) · a9e15c21
  Asaf Joseph Gardin authored Dec 17, 2025
```
Signed-off-by: asafg <39553475+Josephasafg@users.noreply.github.com>
```
  a9e15c21
- Fix nemotron_nas intermediate_size computation (#30795) · f5db6385
  Grzegorz K. Karch authored Dec 17, 2025
```
Signed-off-by: Grzegorz Karch <gkarch@nvidia.com>
```
  f5db6385
16 Dec, 2025 5 commits
- [Core][MM] Optimize encoder cache manager by operating with embeddings only (#30475) · f5f51e59
  Roger Wang authored Dec 16, 2025
```
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Sun Kim <sunytokki@gmail.com>
```
  f5f51e59
- Remove `head_mask` from Ultravox and Swin (#30764) · 0b0acc75
  Harry Mellor authored Dec 16, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  0b0acc75
- Don'e assume `position_embedding_type` will be present for BERT and RoBERTa models (#30770) · 6f15ac5d
  Harry Mellor authored Dec 16, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  6f15ac5d
- [Bugfix] Fix broken ViT attention selection for Blackwell device (#30731) · e94384bb
  Isotr0py authored Dec 16, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  e94384bb
- [CustomOp] Extract ApplyRotaryEmb as CustomOp and unify the dispatch logic (#29873) · 3bd9c491
  Shanshan Shen authored Dec 16, 2025
```
Signed-off-by: shen-shanshan <467638484@qq.com>
Co-authored-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
```
  3bd9c491
15 Dec, 2025 5 commits

[Bugfix] Fix multimodal configuration for Qwen3VL MOE model (#30670) · 3f175f18
Max Hu authored Dec 15, 2025
```
Signed-off-by: Max Hu <hyoung2991@gmail.com>
```
3f175f18
[BugFix] Add embed_input_ids method to make QWenLMHeadModel a vllm model (#30674) · e4806d97
duke authored Dec 15, 2025
```
Signed-off-by: root <iwzbi@zju.edu.cn>
Co-authored-by: root <iwzbi@zju.edu.cn>
```
e4806d97
[Model] Automatic conversion of TokenClassification model (#30666) · 4429d934
wang.yuqi authored Dec 15, 2025
```
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
```
4429d934

[New Model] BAGEL support (AR only) (#28439) · 1adeb3b8

汪志鹏 authored Dec 15, 2025


Signed-off-by: princepride <wangzhipeng628@gmail.com>
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

1adeb3b8

[CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend... · 87b4d155

Shanshan Shen authored Dec 15, 2025


[CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend of QwenVisionAttention with it. (#30125)
Signed-off-by: shen-shanshan <467638484@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: tjtanaa <tunjian.tan@embeddedllm.com>

87b4d155

14 Dec, 2025 4 commits

[Feature]Add EVS (Efficient Video Sampling) Support for Qwen3-VL (#29752) · ae88aada
ZiTian Zhao authored Dec 14, 2025
```
Signed-off-by: zitian.zhao <zitian.zhao@tencentmusic.com>
Co-authored-by: deitxfge <huhaibo1990@126.com>
```
ae88aada
[Bugfix] Revert Qwen2-VL part of change in #28271 (#30542) · 48b8456f
zifeitong authored Dec 14, 2025
```
Signed-off-by: Zifei Tong <zifeitong@gmail.com>
```
48b8456f
[torch.compile] Add encoder tag for compilation (#30489) · 3224ea99
Ilya Markov authored Dec 14, 2025
```
Signed-off-by: ilmarkov <markovilya197@gmail.com>
```
3224ea99

Add AudioFlamingo3 model support (#30539) · 3a20450d

Lasha Koroshinadze authored Dec 14, 2025


Signed-off-by: Lasha <26011196+lashahub@users.noreply.github.com>
Signed-off-by: Lasha Koroshinadze <26011196+lashahub@users.noreply.github.com>
Co-authored-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

3a20450d